«ISSN: 0975-766X CODEN: IJPTFI Available Online through Research Article ADVANCED DETERMINATION OF NODE-MISBEHAVIOUR USING ...»
B.S.Sathish* et al. International Journal Of Pharmacy & Technology
Available Online through Research Article
ADVANCED DETERMINATION OF NODE-MISBEHAVIOUR USING OVERHEARING AND
AUTONOMOUS AGENTS IN WIRELESS AD-HOC NETWORKSB.S.Sathish1*, Dr.P.Thirusakthimurugan2, Ganesan P1, 1V.Kalist 1 Faculty of Electrical and Electronics Engineering, Sathyabama University, Chennai, Tamilnadu, India.
2 Department of EIE, Pondicherry Engineering College, Pondicherry, India.
Email: email@example.com Received on 13-05-2016 Accepted on 12-06-2016 Abstract In Wireless Ad-hoc networks, nodes-operate among themselves to forward information packets from a supply node to a destination node. Nodes might participate in route discovery or route maintenance method however refuse to forward packets owing to presence of faulty hardware or package or to save lots of their resources, such as, battery power and information measure. Detection and isolation of misbehavior nodes area, unit vital problems to boost the standard communication service and to save lots of resources of well behaving wireless nodes, During this work firstly, a neighbour Overhearing based mostly misbehaviour Detection (OMD) theme is planned. In OMD, each node transmissions of its neighbours and calculates packet forwarding quantitative relation of its own in addition as its neighbours. Source node uses the calculated data to spot a misbehaving node. Secondly, associate Autonomous Agent based mostly misbehaviour Detection (AAMD) technique is planned.
Keywords: Wireless Ad-hoc Networks, Misbehaviour, Overhearing, Hash chain, Trusted Third Party, Autonomous Agent, Victimisation, Route discovery
1. Introduction Wireless ad-hoc network is a decentralized network which is not supported by any pre-existing infrastructure. Nodes in this type of network work collaboratively to realize end to end communication. Multi-hop routes are used to overcome limited communication range of nodes in the network. In wireless ad-hoc networks, a source node relies on intermediate nodes to forward data packets to a designated destination node. Wireless ad-hoc network is deployed in some hostile or uncontrolled environment where nodes may not behave according to the defined protocol. So, there will be a probability IJPT| June-2016 | Vol. 8 | Issue No.2 | 13209-13218
and other resources, may provide false routing information or may drop packets to degrade the performance of the network. Misbehaving nodes can be of different types, an overloaded node may lack of processing power, buffer space or available network bandwidth to forward packets. A faulty node might have a software fault where as a malicious node may launch a DOS attack. A selfish node may refuse to forward packets to save its battery power and other resources to forward its own traffic. Existing solutions for identifying misbehaving nodes either require some type of hardware support or use some form of per-packet evaluation of peer behaviour. Credit based approaches require some type of hardware support or a payment system and do not identify misbehaving node. Reputation based approaches are based on transmission overhearing which are expensive and incur more communication overhead. Acknowledgment based approaches require issuance of per-packet acknowledgment, thus introducing high communication overhead. Audit based approach introduces high communication overhead and identification delay for route discovery and audit process.
Moreover, existing techniques fail to detect dropping attacks accurately and efficiently. The rest of the paper is organized as follows. In section II a brief details of proposed work compare with existing work on misbehaviour detection is given.
Section III discusses the concepts related to proposed mechanisms. In section IV the experimental setup and simulation results are shown. Concluding remarks are presented in Section V.
2. Related Work Various techniques have been developed to detect the dropping misbehaviours in wireless ad hoc network. Proposed
schemes can be classified as:
A. Credit Based Systems In credit based systems, some credits are given to intermediate nodes whenever a node forwards packets or provide services to other nodes in the network. The credits received can be used by the nodes to transmit their own traffic. Buttyan and Hubaux have proposed two models which can be used in credit based schemes: Firstly, Packet Trade Model in which each intermediate node gives some credits to previous node and buy packets from it and sells the packets to next node for more credits.
The destination node bears the overall cost of forwarding the packets. Secondly, Packet Purse Model in which credits are loaded into packet before it is sent by the sender node.
Reputation of a node is calculated by its neighbour by observing the packet forwarding behaviour of the node. Reputation computed is used in the system for evaluating the trustworthiness of nodes in forwarding traffic and detecting the misbehaviour of suspicious nodes. This in sequence is then propagating all the way through the network so that the detected misbehaving node can be removed from the network. Marti et al. have proposed a system that contains two most important modules, term watchdog and pathrater, to detect and mitigate routing misbehaviour in MANET, correspondingly. Using the declaration of watchdog module, pathrater rates each path in the cache and selects the path which avoids node misbehaviour.
C. Acknowledgment Based Systems Acknowledgment based systems rely on the reception of acknowledgment to verify that a message is forwarded to next hop. Balakrishnan et al. have proposed TWOACK scheme in which a node sends 2-hop acknowledgment message, whenever they receive a packet, along the reverse path which verifies that the intermediate node has forwarded the packet.
Samreen et al. have proposed an approach which uses two techniques: 2ACK technique and Principle of Flow of Conservation (PFC) technique which are used in parallel to enable the detection of nodes that exhibit packet dropping behaviour. The result generated by first technique is used by the second technique to generate the list of misbehaving nodes.
D. Audit Based System Audit based Misbehaviour Detection (AMD) technique integrates three modules:(a) Reputation module is responsible for management of reputation value of nodes in the network and for updating the reputation based on the input provided by the audit module.(b) Route discovery Module is responsible for finding most trustworthy path among available paths on the basis of path reputation value and path selection factor.(c) Audit module identifies the misbehaving nodes in the path using auditing process. This process is accelerated based on the reputation values provided by the reputation module.
AMD evaluates node misbehaviour on per packet basis without using overhearing technique or acknowledgment scheme.
AMD can also detect dropping attack of selective nature. In this work, firstly, a neighbour overhearing based misbehaviour detection scheme is proposed. Secondly, an autonomous agent based misbehaviour detection scheme is
In this section, firstly, the system model and basic definitions are bestowed. Secondly, projected neighbour Overhearing primarily based misdeed Detection (OMD) technique has been mentioned within which every node gift within the path between supply node and destination node calculates Packet Forwarding Ratio (PFR) of its own and its neighbours.
Supply node maintains a matrix that is updated on the premise of the PFR values calculated by the nodes. This matrix is employed to spot misbehaving nodes. Thirdly, projected Autonomous Agent primarily based misdeed Detection (AAMD) technique has been bestowed within which associate degree agent resides at every node which is able to be activated to calculate PFR of that node as shown in figure 1.
A. System Model The term neighbour is employed to seek advice from a node that's at intervals wireless transmission varies of another node. Neighbourhood refers to any or all the nodes that area unit at intervals wireless transmission varies of that node. A hash chain may be a sequence of hash prices that area unit computed by iteratively line of work a unidirectional hash performs on associate degree initial value. A crucial property of the hash chain is that its parts will be simply computed in one direction, however not within the reverse direction as shown in figure 2, Table 1.
B. Neighbour Overhearing primarily based misdeed Detection: a light-weight approach In this work, the projected neighbour Overhearing primarily based misdeed Detection(OMD) technique uses the data provided by the neighbours for behaviour analysis of nodes within the path while not acquisition high communication overhead. Associate degree algorithmic program for overhearing primarily based misdeed detection is shown in (Sath
matrix contains node id denoted as i, forwarding quantitative relation calculated by its own denoted as PFR and forwarding quantitative relation calculated by its neighbors denoted as PFR wherever j denotes all the neighbours of node metal.
Sath Node Algorithm Rule 1: Algorithm for Overhearing Based Detection of Misbehaving Node.
Require: Forwarding ratio PFR of each node, time epoch t and t1.
Ensure: Detection of misbehaving node.
Initialize matrix φ PFR of Packet Forwarding Ratio (PFR) with zeros.
Selection of path PDS−D using DSR algorithm between source node S and destination 2.
Monitoring packet delivery ratio η of the path PS−D by node S.
If η η0 4.
Node S sends a control packet Request PFR to request all the nodes along the path to calculate PFR of its own 5.
PFRii and its neighbours PFRi (where j denotes all the neighbours of node i) at time t for a time duration t1.
Calculation of PFRii and PFRij by the intermediate node.
At time t+t1 node D sends a reply Request PFR along the reverse path and all intermediate nodes append the 7.
calculated values to the reply.
Node S modifies it’s matrix φPFR based on the received Request PFR 8.
Node S checks all the entries 9.
If PFRii ≠ PFRij 10.
Node i is misbehaving node.
Node i is well behaving node.
Fig. 2. Calculation of PFR values by nodes using neighbour overhearing and verification of misbehavior.
1 0.5 0.5 0.5 2 0.6 0.6 0.6 3 0.6 0.4 0.4 4 0.4 0.2 0.4
4. Simulation Results In the simulation, packet dropping has been implemented under channel conditions given in JProwler. The performance metrics such as Communication Overhead, Identification Delay and Packet Delivery Ratio have been used to evaluate the performance of the proposed scheme. Figure 3 shows the communication overhead introduced to detect misbehaving nodes with respect to percentage of misbehaving nodes. The Y axis is in logarithmic scale. If number of misbehaving nodes increases in the network, communication overhead to detect those node will also increases as more number of agents that are residing at the nodes need to be activated.
Fig. 3. The comparison of communication overhead between AMD and AAMD v/s percentage of misbehaving nodes The control packets need to be transmitted to more number of nodes to collect the calculated PFR values. In OMD, increment in identification delay is less as the number of misbehaving nodes increase because a single control packet can be used for detection of more misbehaving nodes. Identification delay will increase in AAMD in the presence of more misbehaving nodes because more number of agents will be activated for detection. Figure 4 shows the identification delay
7 6 5 4
7 6.5 6 5.5 5 4.5
5. Conclusion We have proposed two techniques for misbehaviour detection in wireless ad-hoc networks. Firstly, an Overhearing based Misbehavior Detection (OMD) technique is proposed which is based on behaviour evaluation by neighbouring nodes. The OMD technique reduces the identification delay by transmitting few control packets for identifying the misbehaving nodes in the network. Secondly, an Autonomous Agent based Misbehavior Detection (AAMD) technique is proposed. The AAMD technique achieves up-to 25% less communication overhead than the AMD technique. Also, the identification
network. Further, investigation may be done to develop an isolation scheme to isolate the detected misbehaving nodes from the network.
References A. Visconti and H. Tahayori, “Detecting misbehaving nodes in manet with an artificial immune system based on 1.
type-2 fuzzy sets,” International Conference for Internet Technology and Secured Transactions, ICITST’09, pp. 1–2, 2009.
2. H. Luo, P. Zerfos, J. Kong, S. Lu, and L. Zhang, “Self-securing ad-hoc wireless networks,” in In Proc. of ISCC, 2002.
3. R. R. Rout and S. K. Ghosh, “Enhancement of lifetime using duty cycle and network coding in wireless sensor networks,” IEEE Transaction on Wireless Communication, vol. 12, no. 2, pp. 656–667, 2013.
4. B. Kannhavong, H. Nakayama, Y. Nemoto, N. Kato, and A. Jamalipour, “A survey of routing attacks in mobile ad hoc networks,” IEEE Wireless Communications, vol. 14, no. 5, pp. 85–91, 2007.
5. A. B. Abderrahmane Baadache, “Fighting against packet dropping misbehaviour in multi-hop wireless ad hoc networks,” J. Network and Computer Applications, vol. 35, no. 3, pp. 1130–1139, 2012.
6. D. Djenouri and N. Badache, “On eliminating packet droppers in manet: A modular solution,” Ad Hoc Networks, vol.
7, no. 6, pp. 1243–1258, 2009.
7. F. Kargl, a. Klenk, M. Weber, and S. Schlott, “Advanced detection of selfish or malicious nodes in ad hoc networks,” in 1st European Workshop on Security in Ad-hoc and Sensor Networks, Heidelberg, Germany, Aug 5-6, 2004, 2004.